Biology Reference
In-Depth Information
35 nucleotides before and after the start codon as the input. This is predicated on a basic
model of translational initiation where the mRNA unfolds as the ribosome initiation complex
is created. In this model, RBS strength depends on the energy required to unfold the sequence
upon translational initiation, the energy released from hybridization of the mRNA to the 16S
rRNA, the energy released upon initiating tRNA hybridization, the energy of unfolding the
standby site, and the energy cost of suboptimal spacing to the start codon. To compute the
first four energy terms, standard RNA folding algorithms were used; to compute the last term,
an empirically determined relationship was used. The result is therefore an RBS part generator
that encompasses a large range of context variations in the several energy terms of its model.
As discussed above, the RBS calculator essentially reduces the library that must be screened to
hit a desired translational efficiency. However, in this case, UTRs and part of the sequence of
the gene of interest must be recoded in each context, and the sequence to activity relationship
is complex
to achieve a
particular activity, such as accomplished in multiplex automated genome engineering (MAGE,
see below), more difficult, and reuse of parts more costly.
something that makes online methods of
'
sequence tuning
'
In a similar vein, Keasling and colleagues used a model-driven approach to create a family
of RNA-regulated devices to control gene expression through mRNA degradation in E. coli ,
in which ribozyme cleavage in a transcript
s5
-UTR leads to a 5
OH-RNA that is more
'
'
'
PPP-RNA. 7 The authors describe an expression control unit
composed of a promoter, a variable ribozyme/aptazyme, and an RBS that can be coupled to
any gene of interest. Performance of this unit is then modeled in great detail using a set of
25 chemical reactions that capture a variety of physical processes that emerge from the
context in which the unit must perform, including transcriptional initiation and elongation,
ribozyme cleavage and folding, and differential mRNA degradation. Some of these variables
are experimentally determined (e.g. ribozyme cleavage rate), while others are inferred from
computational simulations (e.g. ribozyme folding rate). Using this model, they created a
number of ribozyme- and aptazyme-controlled expression systems that hit predicted levels
of gene expression with high precision and integrated these expression control units into a
model metabolic pathway to control the flux of metabolite production. Like the RBS
calculator, this example represents the development of physics-based sequence-activity
models that capture a part
stable than the precursor 5
'
70
s behavior across different contexts through the consideration
of a significant number of variables.
'
Models of these sorts are key to the success of the field. They capture the relevant biophysics
of the parts and, sometimes, their interactions, and suggest places where good design might
eliminate complexities (e.g. the UTR
gene interaction discovered in the ANOVA analysis
above). The simpler the description of sequence-activity can be made by good design
choices, and the more compact the sequence region that encodes a given parameter of the
model, the easier it will be for designers to rationally explore a parameter space to optimize
the behavior of their circuits given the uncertainty in the parts and their operation.
Compositional Insulation
As physical models for predicting the behavior of parts become more complex,
parameterization requires more measurements, which can become excessively
time-consuming if these parameters depend heavily on each different context in which
a part will be used. Therefore, a number of techniques designed to insulate parts from
surrounding contexts have been developed. These techniques mitigate part variability by
removing possible interactions with surrounding contexts, thus reducing the number
of variables that can affect a part
s behavior and simplifying parts-based assembly
of new behavior for all approaches.
'
For example, Davis et al. recently created a set of promoters that contained flanking
sequences around the minimal
50 bp functional promoter in E. coli. 12 The inclusion
B
Search WWH ::




Custom Search